Watch Waveriding or Get Wet?

Does use of artificial intelligence necessarily pose the risk of diminishing critical thinking or thinking skills? The answer might well depend on how AI is used.

But that is true of many human endeavors. People rarely stop engaging in activities they find intrinsically rewarding simply because technology makes the outcome easier to obtain.

As much as I enjoy this clip of waveriding at a favorite spot, I'd much rather be doing it.

Humans experience deep satisfaction when engaged in challenging activities that match their abilities. The reward is not merely the finished product but the experience of making it.

That is why people:

• climb mountains despite vehicle access

• bake bread despite supermarkets

• garden despite grocery stores

• play golf despite television coverage. 

The activity itself provides enjoyment. If people only cared about outcomes, very few would actually play sports.

And AI may affect creative work in much the same way.

Calculators changed mathematics, for example. They reduced arithmetic effort but did not eliminate the need to formulate problems or interpret results. 

There is a reasonable argument to be made that outsourcing “writing” to a language model poses a risk. 

For many writers, that is a relatively negligible risk, since many writers compose because they enjoy the process of writing, and it makes no sense to outsource the “fun” of writing at all. 

For many writers, writing is enjoyable because it combines:

• discovery

• exploration

• self-expression

• problem solving

• craftsmanship.

The point is that technological advances rarely eliminate hobbies.

• Photography did not eliminate painting

• Recorded music did not eliminate amateur musicians

• Power tools did not eliminate woodworking

• GPS did not eliminate hiking

• Word processors did not eliminate writing.

For many forms of knowledge work, AI is not replacing thinking so much as changing where the thinking occurs. 

Instead of spending most of one's effort locating information, more of the cognitive effort shifts to:

• Framing the problem before asking AI

• Evaluating and challenging the response afterward

• Synthesizing the results into an original conclusion.

So, in many research and knowledge-work settings, AI functions less as a replacement for thinking than as an accelerator for information retrieval and synthesis. 

The important intellectual work frequently occurs before the AI interaction (defining the problem, framing the question) and after it (evaluating, integrating, and applying the results).

That is similar to how experienced researchers have long used search engines and databases, for example, and suggests that “how” AI is used matters.

The important question is therefore not "did AI do part of the work?" but " which parts of the thinking did the human still perform?"

Researchers may spend less time collecting information and more time deciding what the information means. They might ask more or different questions.  

And experts might benefit from using AI more than novices, as they are able to formulate better questions, based on:

• relevant mental models

• domain knowledge

• intuition about good questions

• ability to detect errors. 

Stage	Traditional research	AI-assisted research	Where critical thinking occurs
Define question	Formulate hypothesis	Formulate prompt/problem	Very high
Gather information	Library, databases, Google	AI search or LLM	Moderate
Evaluate evidence	Read sources	Verify AI claims and sources	Very high
Compare viewpoints	Read competing authors	Ask AI to generate opposing views	Very high
Draw conclusions	Human synthesis	Human synthesis	Very high
Communicate findings	Human writing	Human writing (possibly AI-assisted editing)	High

Perhaps there is an analogy to the use of “search.” When Google became dominant around 2000, educators raised similar concerns about search making us dumber. 

Instead, search shifted the balance of cognitive work away from memorization and toward higher-level reasoning.

AI might be similar, in many instances.

Thought traditional search asks users to perform much of the information synthesis themselves (find sources and evaluate them), 

AI saves time by summarizing results.

But it does not eliminate the need to ask:

• Is this correct?

• Is something missing?

• What assumptions underlie this answer?

• What evidence contradicts it?

• Are these the strongest sources?

Skeptics will note that many users will not take the time to do so. But that arguably was the case beforehand. 

Rather than replacing reasoning, AI often expands the range of questions that can be explored within a fixed amount of time. 

Also, there are some techniques that encourage broader and deeper exploration.

Recent research describes phenomena such as "cognitive offloading," "epistemic atrophy," and an "illusion of understanding," where users mistake fluent explanations for genuine comprehension. These effects appear strongest when AI substitutes for independent evaluation rather than supporting it. (Business Insider)

For experienced researchers, AI is often best understood as an unusually capable research assistant rather than an autonomous thinker. 

But the researcher still bears responsibility for asking the right questions, verifying sources, weighing evidence, and integrating insights into an original conclusion.

In that sense, AI resembles an evolution of search rather than a replacement for thought. 

The cognitive work shifts away from locating information and toward framing, evaluating, and synthesizing it. 

AI substantially reduces the effort required for search, retrieval, summarization, and drafting, but arguably does not eliminate the need for problem formulation, judgment, skepticism, synthesis, or decision-making.

Whether critical thinking declines depends less on the technology than on whether users treat AI as an answer machine or as a research collaborator whose outputs require evaluation.

For writers who enjoy the process of writing, AI is not a replacement, anymore than watching surfing is a replacement for surfing. 

Data Center NIMBY

A Gallup survey suggests 71 percent of Americans do not want a data center built where they live.

That “not in my backyard” (NIMBY) pattern occurs all the time, concerning landfills, homeless shelters, prisons, mental health facilities, wind turbines, airports, and cell phone towers, for example.

The dynamic reoccurs because of a “mismatch” between benefits and costs. When a policy’s benefits are broad and diffuse but its costs are concentrated, the people who bear the costs have the strongest incentive to organize, while the beneficiaries are often too scattered to mobilize. 

That asymmetry helps explain why many useful policies and facilities trigger intense local resistance even when they are socially valuable overall.

In part, that is because the small number of recipients of the concentrated costs can mobilize easily. The perhaps millions of citizens who benefit are very hard to organize. 

Politically, this is known as a “collective action” problem. The dispersed beneficiaries each have only a small personal stake, while the concentrated losers have a large one, so they show up at hearings, file lawsuits, and pressure officials far more effectively than the hard-to-organize majority. 

So high-performance data centers are not “just” about technology: they also are about politics.

Data-center builders will need to engage in politics to get their projects built. 

Neocloud Depreciation Might Matter, But Perhaps No More Than Supply and Demand

Depreciation schedules  do not often assume strategic importance, but for neocloud suppliers of artificial intelligence “compute as a service,” depreciation of graphics processing units does seem to matter.

 

Assume a $12 billion investment in GPUs. If one assumes a three-year depreciation cycle, that produces a $4 billion per year hit to earnings.

On the other hand, if one assumes a longer six-year cycle, annual depreciation is just $2 billion per year. 

The danger is the balance sheet hit if real-world useful life turns out to be less than six years.

Microsoft, Google, and Amazon arguably can absorb a bad depreciation call because they have robust other sources of revenue.

Neocloud providers must rely almost exclusively on the revenue from their GPU rental businesses. 

So depreciation policy is an existential, not cosmetic issue. A useful-life error doesn't dent one segment's margin, it distorts the entire income statement, since:

• Debt is often GPU-collateralized. Many neocloud financings are underwritten against assumed residual values. Most GPU financing deals assume a uniform, one-size-fits-all depreciation curve. If the real curve is steeper, the collateral coverage on that debt erodes faster than the loan amortizes.

• Contract duration and useful life need to line up. If repayment schedules are structured around a six-year life but revenues fall after three years, debt servicing can become strained.

So schedule length maps to equity valuation. Longer schedules:

• Lower annual depreciation, leading to higher reported net income and earnings per share.

• Improve profitability optics today but risks painful impairments:  if hardware is retired or written down early, the deferred expense hits all at once (an earnings "cliff") rather than being smoothed.

• Skeptics argue that extended depreciation distorts actual operating metrics.

Shorter schedules also affect valuations:

• Cause lower near-term margins and earnings per share, which can produce a valuation discount on trailing/forward P/E versus a peer using a longer schedule, even if the underlying cash economics are identical.

• Lower restatement/impairment risk, and probably a lower cost of capital over time if investors reward accounting conservatism with a quality premium once the market re-prices this issue.

How much does it actually matter? 

Some argue that since depreciation is a non-cash item, "the market sees through it." 

• Free cash flow is identical whether the schedule is three years or six, since the cash left the building at purchase. 

• A DCF-based or FCF-multiple-based valuation should be unaffected by the choice. 

Others argue that generally accepted accounting practices still move equity prices. In this view, GAAP figures can be misleading because they're susceptible to noncash charges like depreciation, and metrics like EPS don't fully reflect a company's profitability. 

So multiple-driven valuation can be distorted by the schedule choice, and equity-linked debt covenants (leverage ratios, EBITDA-based tests) can be gamed as well.

As always, the assumptions matter. Some argue GPU useful life really is in the three-year range. 

Others argue the GPUs still can be used for other operations, and therefore warrants a longer useful life.  

In other words, even when a GPU cannot be used for training, it still has value for inference, other batch work or non-AI operations, generating revenue all the while. 

Physical failure and retirement data for hyperscaler fleets do suggest older GPU generations last seven to nine years in production before physical retirement.

Company	Disclosed useful life	Posture
CoreWeave	~6 years	Most aggressive among pure-play neoclouds
Nebius	3-10 years (blended)	More conservative, cohort-dependent
Lambda Labs	~5 years	Middle ground
Microsoft / Oracle	Extended from 3-4 to ~5-6 years	Matches neocloud aggressiveness
Amazon (AWS)	~4 years	More conservative among hyperscalers
Meta	Up to 11-12 years in places	Outlier on the long end

Either way, the industry seems to be settling on a six-year depreciation cycle for GPU hardware.

Has AI Model Market Begun to Stabilize Around a "Rule of Three" Shape?

At least on mobile devices, ChatGPT remains the share leader, followed by Gemini and then Claude, say analysts at Apptopia. Probably the biggest change is how much share Meta AI has gained, as it now is the fourth-largest model, in terms of mobile device use. 

source: Apptopia 

And it appears the model market is approaching a zero-sum game, where one model’s gains must come from another’s loss, as global usage rates continue to slow. That means a model mostly grows by taking share from another provider. 

And while the market is not yet “stable,” opportunities to dramatically reshape market structure are dwindling. The rule of three appears to be shaping up. 

Capital-intensive industries tend to reach a stable share pattern led by three firms. Just as significantly, the market leader will tend to have twice the share of provider two, which in turn tends to have twice the share of provider three.    

And since share tends to correlate with profit margin, it really matters whether a firm is first or second in a market. Often, the market leader has four times the market share of provider number three. 

By now, if a model is not in the top three, it is very unlikely to break into the top ranks, history might suggest. 

source: Apptopia

AI and Jobs: Correlation is not Causation

It always is difficult to separate correlation from causation in any complex endeavor. Consider the impact artificial intelligence might have on employment. 

Big layoffs at enterprise-sized firms, said to be driven by new AI potential, essentially shift spending from people to tokens but without clear direct financial returns. 

So although we are very early in the process of adopting AI, we still know very little about actual AI impact on jobs. 

A new study by Ramp and Revilio Labs that suggests artificial intelligence adoption actually increases the number of jobs at firms using AI, rather than decreasing employment. 

Or does it?

The study itself suggests a possible “correlation” rather than direct causation: “Companies that adopt AI look very different from companies that never adopt,” the report notes. “AI adopters are larger, more engineering-intensive, more likely to be venture-backed, and were already growing at a faster rate before adoption.”

And that might suggest correlation: the AI adopter firms were growing faster even before AI was adopted. 

It might plausibly also be the case that companies best able to make AI investments can do so because they already are growing revenues and headcount. 

source: Revelio Labs 

“Companies making the largest AI investments grow employment by roughly 10 percent on average following adoption, while low-intensity adopters see no statistically significant change,” the report states. 

Again, the point is that fast-growing firms typically are those adding headcount faster. 

And when the report notes that “among companies making the largest AI investments, the share of entry-level workers increased by 1.15 percentage points compared to not-yet adopters, that might also be because such firms are increasing employment virtually across the board. 

That is not to say AI adoption did not aid employment growth, but only to say we cannot really prove AI was the difference maker, as the data shows the firms adding AI services or apps were faster-growing before AI was added. 

That sort of thinking is in line with other studies of technology adoption that tend to show better-managed firms also are better at integrating new technology. 

Study/Paper	Key Findings	Source
Bloom, Sadun & Van Reenen (2016/2017): "Management as a Technology?"	Management practices (WMS) explain ~30% of TFP gaps; treated as technology-like capital; positive interaction with IT; large cross-country/firm variation.	NBER w22327
ONS (2025): Management practices and technology/AI adoption in UK firms	Strong correlation: better management → higher tech adoption; tech adopters have ~19% higher labor productivity after controls; management predicts AI follow-through.	ONS Article
Cirera et al. (various, e.g., 2021): Firm-Level Technology Adoption (FAT) surveys (Vietnam, Brazil, etc.)	Management quality (incentives, monitoring) strongly predicts technology sophistication indices; linked to productivity; firm capabilities key driver.	World Bank
Babina et al. (2024): AI, firm growth, and product innovation	AI-investing firms show higher sales/employment/valuation growth via innovation; selection via instruments (university AI supply).	ScienceDirect
Alfaro-Serrano et al. (2021): Interventions to promote technology adoption	Reviews evidence linking adoption to performance; management/human capital as key enablers.	PMC
World Bank FAT-related (e.g., Ceará, Senegal)	Management practices and skills correlate with tech adoption intensity; implications for productivity gaps.	World Bank

Better-managed firms might have strong practices in monitoring, incentives, target-setting and talent management, for example. In other words, they have intangible assets that help explain why they are better able to take advantage of new technologies. 

Such firms often also have higher productivity, growth rates and profit margins, making it hard to isolate technology's independent contribution to outcomes. That might be the case with the Revelio Labs study. 

Conversely, poorly-managed firms may lack the complementary skills, processes, or culture to adopt effectively, leading to slower or failed implementations, perhaps with near-term productivity dips as organizational effort is shifted to learning how to use the new tools.

Highly-publicized mass layoffs often are said to be about AI displacement, but often are mostly about correcting earlier overstaffing or simple ways of shifting budgets from people to investing in AI. 

The point is that we cannot discern much, yet, about the actual impact of AI on jobs.